The function of vision is to generate a representation of 3-D objects and surfaces that can be used to achieve current goals (e.g., accurate object identification, efficient navigation). However, two major problems must be confronted in rendering this function. First, the input for vision-the retinal image--is fragmented in space and time. Fragmentation is a consequence of occlusion in 3-D space; an occluding surface may cause two regions belonging to the same object to be separated in space, or a single region to be interrupted in time. Therefore, a critical early operation in vision is to organize image fragments into coherent wholes'; this process of perceptual organization has been a focus of perceptual theorists at least since the Gestalt psychologists began to catalog the principles of organization over 80 years ago. Second, because the brain in general and the visual system in particular are finite in size and capacity, it is not possible to efficiently identify all objects in a visual scene simultaneously. These capacity limitations are surmounted by an attentional mechanism that selects only relevant or salient image regions and thereby reduces the computational complexity of vision. The long-term objective of this research program is to converge on a comprehensive theory of early and intermediate vision, concentrating particularly on how efficient visual selection is accomplished and how it interacts with perceptual organization mechanisms. Several issues arise in this analysis, including the interaction between stimulus-driven attentional capture and goal-driven attentional deployment, the representational basis of visual selection, and the interaction between perceptual organization and attention. The previous award period of the current project began with a series of experiments concerned with the first of these issues [how does top-down control of attention modulate stimulus- driven attentional capture?} and it continued by beginning to examine the representational basis of selection [is attention directed to spatial locations or to preattentively defined perceptual objects?}. The present competing continuation application continues to address these questions and extends the project to address the third issue what is the interaction between perceptual organization and attention? The specific aims of the current project are (a) to explore the role of amodal integration, an organizational mechanism that is analogous to amodal completion, as a fundamental operon in early perception, (b) to examine implications of figural identity assignment are for perception and attention; and (c) to assess the implications of our recent findings for current theories of visual search and attentional capture. The proposed experiments will provide critical evidence concerning the appropriate architecture for early vision, and will contribute to the overall objective of deriving a comprehensive theory of early and intermediate-level vision.